Level Set Based Image Segmentation with Multiple Regions

We address the difficulty of image segmentation methods based on the popular level set framework to handle an arbitrary number of regions. While in the literature some level set techniques are available that can at least deal with a fixed amount of regions greater than two, there is very few work on how to optimise the segmentation also with regard to the number of regions. Based on a variational model, we propose a minimisation strategy that robustly optimises the energy in a level set framework, including the number of regions. Our evaluation shows that very good segmentations are found even in difficult situations.